Validation and Calibration of Conceptual Design Tool SUAVE

Abstract

The aim of this study is to provide a comprehensive validation of the conceptual design tool SUAVE, with respect to real world flight data from an extensive flight database of 737-800NG fleet operations. A validation of the initial SUAVE model fuel burn predictions was performed and compared with the actual data. A surrogate propulsion model was developed using NPSS and integrated into SUAVE, which significantly enhanced the model accuracy. A calibration of the SUAVE aerodynamic model coefficients, the propulsion model idle throttle value, and the take-off and landing drag increments was performed which aimed to minimise the fleet-wide fuel burn error with respect to the actual flight data within the database. The model was calibrated using 26 case flights in a training database, and the final calibrated model performance was measured against three unseen test cases within the flight database. A modest reduction in error was observed for the training and test cases, however non-linear error effects in the presence of headwinds and tailwinds affected the model accuracy and limited the potential of the calibration. Despite these effects, a highly accurate model was developed with an average fuel-flow error of 5% with respect to real-world flight data, providing an accurate baseline model from which confident projections of future carbon reductions can be obtained for new aircraft designs.